Deleted Posts, et tu Brown?

[dpatent]

Interesting that I gave a long and detailed reply to an issue regarding variant-making and it promptly got deleted.  Seems like that has been going around a lot lately.  

Responses expected.  Let\'s see if you two (JB and Alydar) can get through a post -- like I did -- by dealing with the facts, answering the questions and without saying something insulting.

Here is the string:

Author: TGJB
Date:   06-05-02 12:08

Re: 3 – What’s interesting (and what this is all about in the end) is that they don’t use the same variant, which is why they can’t address any questions about variant making without it becoming clear they are hypocrites. On Preakness Day they took 1 3/4 points more off the third than the second (both routes) then had the track getting faster, for example. And they have the grass course getting significantly slower, which is nuts considering both the conditions and the practical ramifications (figures assigned). They have some rules concerning the relationship between independent events, which they apply, except when they don’t. All the “science” is intended to make it look like the figures come from a higher power.

TGJB
 
 
Reply To This Message
 
 
 Re: Moving On
Author: Alydar in California
Date:   06-06-02 05:24

Ragozin wrote that \"all the figures must use the same variant unless rain or a freeze or a thaw changes things.\" Thus, Ragozin could argue that the rain changed things on that day. What\'s strange is if they have their variants moving against the grain: the track getting slower when Patent\'s meteorologist thinks it should be getting faster, for example. This would invalidate their figures, according to Patent.

What I was ridiculing was Patent\'s idiotic claim, made to cover his retreat, that your numbers and Ragozin\'s numbers would be in sync 80-90 percent of the time even if, as Patent believes, Ragozin ordinarily does all the races at the same variant, and your variants \"yo-yo\" up and down.

Now I\'m done with Patent, barring something unusual.
 
 
Reply To This Message
 
 
 Re: Moving On
Author: David G. Patent
Date:   06-06-02 10:42

Alydar,

At Jerry\'s suggestion I took a look at the sheets for the Preakness horses -- over 100 horses with about 1,500 to 2,000 numbers. This is what I observed:

1) Brown\'s and Ragozin appear to have substantially different scales for certain circuits (confirming what JB said)

2) However, within those circuits, the variants are remarkably similar from race to race -- I would say 70-85% of the time the numbers \'agree\' (+ or - a point or so) once you factor in the different variant -- the sheets look very similar, just \'shifted\' over. Brown\'s horses run tighter patterns than Ragozin, but the average deviation (hi-lo) between numbers is not that much different.

Now, the key here is on the margins. My problem with the TG product is that I make my money (when I do make money) finding horses likely to run bad and playing against them. It\'s finding that one or two horses getting bet that look really bad. That only happens a couple times on a card I can\'t do that as well with TG\'s product because horses look a lot more likely to run good. Thus, my selection of product. Others may have a different method.


3) Jerry\'s claim -- and this is the important one -- is that the fact that his horses maintain a tight range proves that his numbers are right. Now let\'s just take that statement on its face. If that were true, then Jerry would have the ultimate proof that his product was superior and he should be able to demonstrate that superiority day-in day-out to the racing world. The big bettors, few if any of whom are just blind dancing partisans like myself, would use TG and not Ragozin. Is that reality? No. Not even close.

Brown\'s claim is false. The tight ranges do not prove that his numbers are right -- see below. Do you believe that statement of his? I don\'t think anybody but he and (perhaps) you buy that one.

I have said from day 1 that neither Ragozin nor TG can \'prove\' anything. The method that each employs rests upon different sets of assumptions and the product that each of us bettors chooses likely reflects our comfort level with the maker\'s underlying assumptions. Agree? We\'ve gone round and round on empirics vs. belief and I don\'t expect you\'ll ever come around on that one.

What I did see, from looking at those Preakness day horses, however, was that Jerry\'s claim that his numbers would collapse (at least I think that\'s his claim) if his variant was wrong on a day is wrong. As long as horses run within the same circuit there will not be a problem (for example, JB has a horse at CT running straight 5s. On Ragozin, she runs 10s-12s). As long as that horse stays at CT, she will continue, I think, to run tight numbers). Even if a horse moves to a ciruit where the scale is different (this is not a very common occurrence -- note that almost all the Del, Pim, Lrl, horses stay in that circuit) -- I\'m now assuming that Ragozin is \'right\' and Jerry is \'wrong\' -- then the only anomaly will be the single race that the horse has run outside the circuit. That won\'t make the horse\'s sheet collapse at all. It will just be a blip on the radar (MDO\'s Preakness number is a good example -- it looks a lot more plausible on Ragozin\'s sheet than on TG\'s). If the horse stays on the new circuit, the numbers will just shift over. Agree? Only when horses bounce around to different circuits with different scales is there a risk that their TG sheet will collapse.

The question is, how many numbers in a row that were \'off\' would it take for a horse\'s sheet to go to hell. Similarly, what % of a horse\'s number would have to be \'wrong\' for the tight ranges to collapse? My point was that it would take at least 3 (and probably 5) \'bad\' numbers in a row before the sheet started looking funny -- the percentages I associated with the 80% and 90% were correct (8 in 1000 and 1 in 1000; significantly less likely for 5 in a row). If not, tell me how that math was wrong.

The second question, what overall percent would need to be wrong? That\'s a question of judgment. I would wager that it is 25% to 35% (for a horse with at least 10 races or so) before the sheet just didn\'t make any sense. What are the odds of that happening? Depends on how many races the horse runs, but you won\'t find many horses given the odds as long as Rag. and TG agree the majority of the time. Do you agree or disagree with this? If not, tell me specifically what is wrong with the reasoning. Don\'t just say it\'s \'idiotic\' or that I\'m \'unreal\'.

Maybe you take issue with the 70-85% assumption. So, what is the percent in your view? Once you take a stand, it\'s easy to estimate what percent of horses would have sheets that \'collapse\' if Brown got the variant wrong.

Brown\'s latest post on the variant indicates that he usually has the day flat or on a slide, so there can\'t be too much difference between he and Ragozin on a race-by-race basis. When I said that Brown\'s variants go up and down like a yo-yo that\'s what I meant, but it does not mean that every day every race they go up and down. My point was the fact that they occassionally do go up and down raises a red flag and I\'d want to know more about why someone reversed direction on their variant (e.g., the Preakness day). Is that unreasonable?

[Alydar in California]

David: As briefly as possible:

 We offered to answer every one of your questions as long as you answered ours. You said no. You said you were afraid your wife would take your baby away from you. Now you are complaining that we don\'t answer your questions. One can\'t make this up, and one can\'t imagine the effrontery required to ask MORE questions, questions based on false and hilarious premises, to boot.

You have changed your percentages, which I called \"idiotic.\" But your new percentages are only slightly less idiotic. If Ragozin\'s variants are steady and JB\'s yo-yo up and down \"ON ANY GIVEN DAY\" (as you wrote) and do all the other crazy things you\'ve said they do, the numbers will not be in sync 70-85 percent of the time. Now you\'re backing off of \"any given day,\" and you\'re even changing the meaning of \"yo-yo,\" for God\'s sake.  

You ignored my first paragraph, which means you believe Ragozin\'s numbers are invalid.

You have changed positions constantly, backed off numerous times, and are now left with this:

\"JB has a horse at CT running straight 5s. On Ragozin, she runs 10s-12s. As long as that horse stays at CT, she will continue, I think, to run tight numbers [on TG]. Only when horses bounce around to different circuits with different scales is there a risk that their TG sheet will collapse.\"

Completely wrong. What matters is the accuracy of the numbers. Let\'s say there is NO circuit changing. If you start out with artificially tight cycles (inaccurate numbers, in other words), horses to whom you gave 5s will frequently lose to horses to whom you gave 7s. When this happens, your cycles will automatically get looser.

I will make it even simpler for you. 5F race. Only sprint of the day. One length equals about one point at this distance. Five-horse field. Here are their figures going into the race:

Horse 1: 5555555555.
Horse 2: 6666666666.
Horse 3: 7777777777.
Horse 4: 8888888888.
Horse 5: 9999999999.

Cycles cannot get any tighter than these. Then they run. Ground loss and weight are even. Horse 1 wins by one length over horse 2, who finishes one length ahead of horse 3, who finishes one length ahead of horse 4, who finishes one length ahead of horse 5.

What should the variant be for this race? Easy, right? A number that gives horse 1 another 5, horse 2 another 6, etc.

Now let\'s say the numbers we had given these horses GOING IN TO today\'s race had been inaccurate. Let\'s say horse 5 had really been running 5s, horse 4 had been running 6s, and so on. The order of finish will be the reverse of the way I showed it above. What\'s your variant now? What happens to the tightness of the cycles?

Now try something less far-fetched: Pretend that all the figures were accurate except the most recent number for horse 5. Say we had it one point slower than it should have been. Say he pairs it and DHs for fourth. What happens now? You throw him out of your projection (giving four other horses the numbers they usually run), and his cycle gets slightly looser.

Inaccurate figures = \"implausible\" results = a loosening of the cycles in order to incorporate the \"implausible\" results. That\'s another way of saying that fast horses tend to beat slow horses and if your numbers don\'t show which horses are faster and by how much, your cycles will get progressively looser.

 Patent: \"We\'ve gone round and round on empirics vs. belief and I don\'t expect you\'ll ever come around on that one.\"

I think you mean \"empiricism,\" David. In truth, TG and Rags numbers are based on a combination of experience/observation and theory. (By the way, who makes Ragozin\'s Maryland numbers? Who makes his California numbers? Who makes his Illinois numbers? What are the names of these trackmen who are counting the raindrops and estimating evaporation? In your system, they become extremely important.)

Patent: \"I would wager that [25-35 percent of the numbers have to be inaccurate] before the sheet just didn\'t make any sense.\"

You are changing the terms of the debate, which is not a bad idea when your previous position is indefensible, as yours was. The question is whether tight cycles are sustainable if the numbers are inaccurate.

\"What percentage of the horse\'s numbers would have to be wrong for the tight ranges to collapse?\"

See above. One bad number can loosen the cycle, and there is a cumulative effect. Bad numbers do damage in future races, producing more bad numbers and, eventually, chaos: the opposite of the pairs, trios, and pretty numbers you detest.

[Anonymous User]

You\'ve got to try and make it scientific, though I agree with T-Graph about the viability of a racing surface. Still the danger is giving a horse a number because he ran it before. I still disagree with the numbers Commendable earned. I thought his win was a freakish day and surface result. Second by seven lengths to Tiznow, he still earned another two. Which I could have sworn was a three then. The bottom line is you\'ll never get horses to behave as nicely as the projections in hypotheticals and you shouldn\'t try.

[superfreakicus]

nt, damnit!!

[TGJB]

David Patent wrote:
>
> Interesting that I gave a long and detailed reply to an issue
> regarding variant-making and it promptly got deleted.  Seems
> like that has been going around a lot lately.
>
> Responses expected.  Let\'s see if you two (JB and Alydar) can
> get through a post -- like I did -- by dealing with the
> facts, answering the questions and without saying something
> insulting.
>
> Here is the string:
>
> Author: TGJB
> Date:   06-05-02 12:08
>
> Re: 3 – What’s interesting (and what this is all
> about in the end) is that they don’t use the same
> variant, which is why they can’t address any questions
> about variant making without it becoming clear they are
> hypocrites. On Preakness Day they took 1 3/4 points more off
> the third than the second (both routes) then had the track
> getting faster, for example. And they have the grass course
> getting significantly slower, which is nuts considering both
> the conditions and the practical ramifications (figures
> assigned). They have some rules concerning the relationship
> between independent events, which they apply, except when
> they don’t. All the “science” is intended
> to make it look like the figures come from a higher power.
>
> TGJB
>  
>  
> Reply To This Message
>  
>  
>  Re: Moving On
> Author: Alydar in California
> Date:   06-06-02 05:24
>
> Ragozin wrote that \"all the figures must use the same variant
> unless rain or a freeze or a thaw changes things.\" Thus,
> Ragozin could argue that the rain changed things on that day.
> What\'s strange is if they have their variants moving against
> the grain: the track getting slower when Patent\'s
> meteorologist thinks it should be getting faster, for
> example. This would invalidate their figures, according to
> Patent.
>
> What I was ridiculing was Patent\'s idiotic claim, made to
> cover his retreat, that your numbers and Ragozin\'s numbers
> would be in sync 80-90 percent of the time even if, as Patent
> believes, Ragozin ordinarily does all the races at the same
> variant, and your variants \"yo-yo\" up and down.
>
> Now I\'m done with Patent, barring something unusual.
>  
>  
> Reply To This Message
>  
>  
>  Re: Moving On
> Author: David G. Patent
> Date:   06-06-02 10:42
>
> Alydar,
>
> At Jerry\'s suggestion I took a look at the sheets for the
> Preakness horses -- over 100 horses with about 1,500 to 2,000
> numbers. This is what I observed:
>
> 1) Brown\'s and Ragozin appear to have substantially different
> scales for certain circuits (confirming what JB said)
>
> 2) However, within those circuits, the variants are
> remarkably similar from race to race -- I would say 70-85% of
> the time the numbers \'agree\' (+ or - a point or so) once you
> factor in the different variant -- the sheets look very
> similar, just \'shifted\' over. Brown\'s horses run tighter
> patterns than Ragozin, but the average deviation (hi-lo)
> between numbers is not that much different.
>
> Now, the key here is on the margins. My problem with the TG
> product is that I make my money (when I do make money)
> finding horses likely to run bad and playing against them.
> It\'s finding that one or two horses getting bet that look
> really bad. That only happens a couple times on a card I
> can\'t do that as well with TG\'s product because horses look a
> lot more likely to run good. Thus, my selection of product.
> Others may have a different method.
>
>
> 3) Jerry\'s claim -- and this is the important one -- is that
> the fact that his horses maintain a tight range proves that
> his numbers are right. Now let\'s just take that statement on
> its face. If that were true, then Jerry would have the
> ultimate proof that his product was superior and he should be
> able to demonstrate that superiority day-in day-out to the
> racing world. The big bettors, few if any of whom are just
> blind dancing partisans like myself, would use TG and not
> Ragozin. Is that reality? No. Not even close.
>
> Brown\'s claim is false. The tight ranges do not prove that
> his numbers are right -- see below. Do you believe that
> statement of his? I don\'t think anybody but he and (perhaps)
> you buy that one.
>
> I have said from day 1 that neither Ragozin nor TG can
> \'prove\' anything. The method that each employs rests upon
> different sets of assumptions and the product that each of us
> bettors chooses likely reflects our comfort level with the
> maker\'s underlying assumptions. Agree? We\'ve gone round and
> round on empirics vs. belief and I don\'t expect you\'ll ever
> come around on that one.
>
> What I did see, from looking at those Preakness day horses,
> however, was that Jerry\'s claim that his numbers would
> collapse (at least I think that\'s his claim) if his variant
> was wrong on a day is wrong. As long as horses run within the
> same circuit there will not be a problem (for example, JB has
> a horse at CT running straight 5s. On Ragozin, she runs
> 10s-12s). As long as that horse stays at CT, she will
> continue, I think, to run tight numbers). Even if a horse
> moves to a ciruit where the scale is different (this is not a
> very common occurrence -- note that almost all the Del, Pim,
> Lrl, horses stay in that circuit) -- I\'m now assuming that
> Ragozin is \'right\' and Jerry is \'wrong\' -- then the only
> anomaly will be the single race that the horse has run
> outside the circuit. That won\'t make the horse\'s sheet
> collapse at all. It will just be a blip on the radar (MDO\'s
> Preakness number is a good example -- it looks a lot more
> plausible on Ragozin\'s sheet than on TG\'s). If the horse
> stays on the new circuit, the numbers will just shift over.
> Agree? Only when horses bounce around to different circuits
> with different scales is there a risk that their TG sheet
> will collapse.
>
> The question is, how many numbers in a row that were \'off\'
> would it take for a horse\'s sheet to go to hell. Similarly,
> what % of a horse\'s number would have to be \'wrong\' for the
> tight ranges to collapse? My point was that it would take at
> least 3 (and probably 5) \'bad\' numbers in a row before the
> sheet started looking funny -- the percentages I associated
> with the 80% and 90% were correct (8 in 1000 and 1 in 1000;
> significantly less likely for 5 in a row). If not, tell me
> how that math was wrong.
>
> The second question, what overall percent would need to be
> wrong? That\'s a question of judgment. I would wager that it
> is 25% to 35% (for a horse with at least 10 races or so)
> before the sheet just didn\'t make any sense. What are the
> odds of that happening? Depends on how many races the horse
> runs, but you won\'t find many horses given the odds as long
> as Rag. and TG agree the majority of the time. Do you agree
> or disagree with this? If not, tell me specifically what is
> wrong with the reasoning. Don\'t just say it\'s \'idiotic\' or
> that I\'m \'unreal\'.
>
> Maybe you take issue with the 70-85% assumption. So, what is
> the percent in your view? Once you take a stand, it\'s easy to
> estimate what percent of horses would have sheets that
> \'collapse\' if Brown got the variant wrong.
>
> Brown\'s latest post on the variant indicates that he usually
> has the day flat or on a slide, so there can\'t be too much
> difference between he and Ragozin on a race-by-race basis.
> When I said that Brown\'s variants go up and down like a yo-yo
> that\'s what I meant, but it does not mean that every day
> every race they go up and down. My point was the fact that
> they occassionally do go up and down raises a red flag and
> I\'d want to know more about why someone reversed direction on
> their variant (e.g., the Preakness day). Is that unreasonable?

TG---You know bwtter than to think I deleted you on purpose. Much of what you say was dealt with by Alydar (and he\'s right, your premised are ____ed), and lots of this was dealt with by me in earlier posts that you didn\'t pay attention to, ignored, evaded, etc. I\'ll touch on a few points.

Your 1. Hallelujah. See Magic Weisner and Quidnaskra for practical demonstrations, as I said before.

2. The overall sheet may look similar, but I seriously doubt your 70-85% figure. The route/sprint variability is too high, and when ragozin uses an average he has to have at least one of them wrong, or both wrong if he fudges. Examples abound (Quixotes Hope).

3. You actually might not be as smart as I thought.

3a. Who the big bettors use can be a function of many things--hype, lies, peer pressure, and a lack of understanding of the essentials--as in your case. YOU CAN\'T ARTIFICIALLY MAKE THE NUMBERS COME OUT TIGHT EVEN WITHIN A CIRCUIT, and with a few exceptions, race meets feature horses coming from all over.

3b. It\'s not a question of a sheet looking wrong. You simply can\'t have the Peter Pan come out the way it did, or the day to day races, unless the pieces fit, and you can\'t have the pieces fit unless either you mess with the data (earlier figures, relationships within the race), or it\'s right.

3c. Friedman liked MDO in the Preakness, I didn\'t. You\'re going to have a hard time convincing anyone that his running bad works against me. I also don\'t claim that horses run the same number forever, or all horses always run in a tight range--see not only MDO, but WE\'s first big number.

3d. Alydar did a pretty good job with his example, but his earlier statement is dead on--you can\'t have made projection figures or you wouldn\'t be saying any of this. Do it for a week and you\'ll see. Or, just do what you evaded before--show me how I artificially made the Peter Pan numbers come out that way. Specifically, not generally.

I\'m going to post a question on a separate string you might want to think about.

[dpatent1]

Alydar/JB

I think we\'redone with this one.  

Too bad you guys can\'t just debate the merits and feel the need to fire personal insults in virtually every post.  That\'s usually not a good sign for the merits of your argument.

The thing that takes out your response is your penchant for gross oversimplification and makes it impossible to have an intelligent discussion.

This is what you wrote:  \"Now let\'s say the numbers we had given these horses GOING IN TO today\'s race had been inaccurate. Let\'s say horse 5 had really been running 5s, horse 4 had been running 6s, and so on. The order of finish will be the reverse of the way I showed it above. What\'s your variant now? What happens to the tightness of the cycles?\"

Like I said, there is no such thing as \'the numbers being inaccurate.\'  Your example assumes 100% inaccuracy for every horse.  Your other example wasn\'t much better.  My point from the beginning has been that on the margins TG is less useful because he manipulates the variant to achieve a tighter range of results.  My estimate is that this happens about 20% of the time.  I verified that by looking at several hundred results.  No answer from you on this.  Or on whether you actually believe JB\'s argument that a tight range of numbers proves that he is right and Ragozin is wrong.

Having 20% of your numbers be off a few points (assuming Ragozin is right 100% of the time, which he is not) does not cause a collapse of the sheet.  It only causes an occasional anomaly.

Making figures is I agree a matter of judgment and interpretation.  TG has claimed a lot more though -- \"I\'m always right because my cycles are tight\"

In my view the TG bias toward forcing results to be \'tight\' is why TG is not \'wrong\' per se, but less useful.  I have never claimed more than this point.  Look back at my first post on May 23, which, by the way, was a response to JB\'s boast that Rag. got a bunch of races wrong on Preakness day and later had to backpedal like crazy.

[Alydar in California]

Patent wrote: \"Making figures is I agree a matter of judgment and interpretation. TG has claimed a lot more though--\'I\'m always right because my cycles are tight.\'\"

David: Show me where JB wrote that. If you can\'t, tell me why you made it up.

Give me the PRECISE adjustments you used to bring the various circuits into line when you went over the Preakness day figures. Your percentages are not very important to this argument, but I am curious about them. I looked at part of a Hollywood Park day for which I have TG and Rags figures. That\'s why I am interested, very interested, in your adjustments.

I will reply to the rest of this late Saturday night--whether you have replied or not. I will reply to Jason then, too.

[Alydar in California]

Jason: Sorry. I\'ve been out of sorts. I\'ll give you a long reply tonight or tomorrow.

David: Did you miss this?

Nunzio: Congratulations on winning the contest.

HP: Congratulations on beating Patent.

[Alydar in California]

Oh, and David: response expected, as you say.